CTC: Why We Do It
Joseph T. Ferrucci Question: Why do we climb Mt. Everest?
Answer: Because it is there.
Anonymous
J. T. Ferrucci, MD
Chair Emeritus and Professor of Radiology, Boston University School of Medicine, Boston Medical Center, 88 East Newton Street, Boston, MA 02118
that the sensitivity of CTC equaled that of conven- tional colonoscopy for detection of large polyps and cancers in a landmark 100 patient Boston Univer- sity study published in the New England Journal of Medicine (Fenlon et al. 1999). As they say, the rest is history.
As CTC enters its second decade, it is no longer new, but retains many compelling features. Techno- logically it maintains its sophisticated, innovative appeal and still exhibits great potential to evolve further. Scientifically, CTC is reframing strategies for colorectal cancer screening and now challenges the primacy of colonoscopy and the specialty of gas- troenterology for the diagnosis of colon disorders.
At the same time, CTC has been a dominant focus of research in abdominal and gastrointestinal radi- ology for several years, stimulating an enormous volume of original scientific investigation as well as media and industry attention. Impressive clini- cal results continue to appear from investigators throughout the world, including North America, Europe, and Australia (Yee et al. 2001; Macari et al. 2002; Iannacone et al. 2003; Edwards et al.
2004). Even more important is the totally non-inva- sive aspect of CTC (no drugs, no contrast media and no injections) which has won the favor of many physicians and their patients, especially when com- pared to optical colonoscopy. In preference studies comparing the two tests, patients usually prefer CTC despite the unavoidable biases of pre-endoscopy sedation (Svensson 2002). It is this patient friendly,
‘compliance enhancer’ nature of CTC which has been able to attract otherwise reluctant patients to undergo colorectal cancer screening. A recent U.S. hospital think tank reported that some 60%
of patients having virtual colonoscopy had never had any prior form of colorectal cancer screen- ing (Advisory Board 2004) (Fig. 1). In the United States, several HMOs have begun to reimburse for colorectal cancer screening using CTC, and wider reimbursement coverage is expected in 2006 which should lead to rapid wide dissemination into clinical practice.
CONTENTS
1 Introduction 1
2 Colorectal Cancer Screening (CRCS): Rationale 2 3 Colon Polyp: Natural History/Target of Screening 2 4 Clinical Results 4
5 Acceptance of CTC 4 6 Conclusion 5 References 5
1
Introduction
Computed tomographic colonography (CTC), com- monly known as Virtual Colonoscopy (VC) has recently emerged as a fundamentally new tech- nique for radiologic imaging of the colon with the unique potential for broad application in population screening for colorectal cancer. Yet, when framed in the philosophic question of “why do we do CTC?”, the analogy to Mt. Everest becomes clear. We do CTC because the technology exists.
In the early 1990s, the introduction of spiral CT
scanners, and powerful new computer workstations
for image processing prompted individual pio-
neers to exploit the new technology at least in part,
because they could. Coin obtained a United States
patent for CT reconstruction of the colon (Coin et
al. 1995), while Vining is credited with the first clini-
cal demonstration of what he termed ‘virtual colo-
noscopy’ (Vining and Gelfand 1994). Hara at the
Mayo Clinic (Hara et al. 1996) and Royster at Boston
University (Royster et al. 1997) confirmed clinical
feasibility for polyp detection. Fenlon then showed
2
Colorectal Cancer Screening (CRCS):
Rationale
Across the developed world, colorectal cancer is the second or third leading cause of cancer deaths.
While a small percentage (10–20%) of colorectal cancers occur in high risk genetically predisposed patients, the majority, i.e., ca. 80% of colorectal cancers occur sporadically in otherwise low risk individuals. In the vast majority of such cases, the cancers are believed to arise from pre-existing ade- nomatous polyp pre-cursors in series of events that have a well characterized origin in genetic muta- tions with a consequent histopathologic sequence of degeneration into frank invasive cancer. How- ever, this process is rather leisurely, requiring some 10–15 years or more and interruption of this pro- gression by detection and removal of threatening pre-cursor adenomas by endoscopic polypectomy results in a decline of cancer related mortality by as much as 30%.
Guidelines for colorectal screening in asymp- tomatic populations have been developed on the basis of scientific medical evidence, by professional organizations and government agencies throughout Europe and North America. Most recommend that screening begin in asymptomatic individuals at low or average risk at age 50 years and permit several different testing strategies. These include annual screening with fecal occult blood tests, flexible sig- moidoscopy every five years, the combination of fecal occult blood and flexible sigmoidoscopy every five years, double contrast barium enema every five years or colonoscopy every ten years. However, none of these test strategies is ideal and proponents of the various strategies continue to engage in conten-
tious debate. For example, the specific limitations of FOBT and flexible sigmoidoscopy have led to the concept of the desirability of an anatomic or struc- tural examination of the whole colon. This has led to the emergence of colonoscopy as the de facto gold standard for colorectal screening as well as colon diagnosis generally. The more focused debate has revolved around whether colonoscopy should be offered as a universal once in a lifetime test, e.g., at age 60 or reserved for selective application when results of other preliminary screening tests are posi- tive. In the latter case, the broader goal of colorectal cancer screening becomes the use of less invasive, less expensive tests for triage selection of patients to undergo therapeutic optical colonoscopy. (Par- enthetically, the double contrast barium enema is rapidly falling out of favor in the U.S. for primary colorectal cancer screening.)
Yet, despite wide medical, public health and lay media airing as to the importance of colorectal cancer screening, the public has remained generally reluctant to undergo these tests which are perceived as unpleasant and embarrassing such that overall compliance with colorectal cancer screening rarely exceeds 30–40%. Recently in the specific instance of colonoscopy in the United States, compliance rate have increased slightly to ca.40–50%, but only in selected well insured patient groups. Moreover, manpower resources of colonoscopists are strained, at least in the United States, with long 6–12 month waiting lists for elective appointments. Thus, new alternative tests for colorectal cancer screening are needed and awaited and, along with fecal DNA testing, CTC appears a procedure whose time has come.
3
Colon Polyp:
Natural History/Target of Screening
The progressive transformation of adenomatous polyps to invasive adenocarcinoma has been char- acterized as “the adenoma carcinoma sequence”
(Muto 1975). However, because the prevalence of undetected cancer in an asymptomatic screen- ing population is very low at ca.1%, colon polyp size is widely accepted as a surrogate end point for outcomes assessment in colorectal cancer screen- ing programs. Thus, the concept of the “advanced adenoma” has been developed which is defined as an adenomatous polyp measuring 10 mm or greater
Fig. 1. CTC as a ‘compliance enhancer’. Pie chart from a US hospital think tank study showing that among patients having virtual colonoscopy 60% had never had any prior colon cancer screening
VC patients: prior experience with CRC Screening Compliance enhancer
The Advisory Board - Washington, DC June 2004
No Yes
or one containing villous or dysplastic components at histologic examination (Winawer and Zauber 2002). However, in the context of CRC screening, the actual prevalence of advanced adenomas or at least polyps 10 mm or greater is rather low at approxi- mately 5–10%. By the same token, approximately 50% of adults age 50 years will harbor some form of a polyp at colonoscopy with the prevalence of polyps increasing linearly in the 6th and 7th decades there- after. Thus, as Bartrum has stated, it is “normal to have a polyp at age 50 years” (Bartrum 2000).
Herein, the emergence of CTC creates a new conundrum for those assessing colorectal cancer screening strategies. Because colon polyps are so common, while advanced adenomas are relatively rare and frank cancers even rarer still, there is a sudden new focus of interest in better characterizing the natural history of polyps detected by radiologic CTC colon screening. When using optical colonos- copy for screening, every polyp detected is simply removed and it is more or less academic to deliberate further on their histopathology or malignant poten- tial. However, many if not most small polyps are either not adenomas and are merely hyperplasic or non-specific inflammatory response on histology.
Further, even if they are in fact adenomas, they are so small that their potential for malignant degen- eration will never be expressed in the patient’s life (Bond 2001). In studies of colonoscopic screening of average risk populations, approximately 60% of patients had no evidence of neoplasia while another 20–25% of patients had only one or two subcenti- meter polyps (Leiberman et al. 2000; Pickhardt et al. 2003). Thus, in terms of colon cancer preven- tion, the detection and removal of such diminutive polyps will convey little or no direct benefit to the individual patient in terms of cancer prevention.
Considerations such as the above contain epi- demiologic and public health policy implications as they highlight the distinction in benefits from colorectal screening that may apply to a single indi- vidual patient vs the strategic allocation of resources for colon cancer prevention in a population at large.
In this context, the benefit harms ratio of screen- ing asymptomatic populations is a critical concern.
In colon screening, unlike screening for breast or lung cancer, the target lesion for detection is merely a benign precursor, rather than an actual frank his- tological malignancy and the intervention becomes cancer prevention rather than cancer detection.
Thus, it is even more critical to obtain a careful bal- ance of variables such as risk, resource cost and test- ing interval (Fig. 2).
As suggested in the foregoing, the new issue for CRC screening will become the proper manage- ment of polypoid lesions detected at CTC in terms of follow up surveillance or referral to colonoscopy for polypectomy. Guidelines currently recommended by the Working Group for Virtual Colonscopy (a informal federation of VC researchers) is shown in Fig. 3. Of note, is the relative unimportance of polyps less than 5 mm in diameter and the recom- mendation that lesions 10 mm or larger be referred for polypectomy. Much of the future debate, there- fore, will be reserved for management of interme- diate sized polyps 6–9 mm in size. For this group of patients, multiple factors (age, co-morbidities) will require consideration leading to individualized patient decisions.
Fig. 2. Balancing variables in strategies for CRC screening CRC Screening Strategy: Variables
Polyp Target Size
Cost Interval R I S K
RECOMMENDATIONS Colonoscopy ?
% of Patients Recomendations No polyps (c. 40%) NO ACTION
(Routine surveillance)
Polyp 5 mm or < (c. 30%) NO ACTION
(Routine surveillance)
Polyp 6–9 mm (c. 20%) INDIVIDUALIZE pt age, co-morbidities ? Multiple polyps esp > 3 Polyp 10 mm or > (c. 10%) COLONOSCOPY
Fig. 3. Recommendations for management of CTC detected polyps developed by the Working Group for Virtual Colonos- copy (WGVC), at the 5th International Symposium for Virtual Colonoscopy, Boston Oct 2004
4
Clinical Results
CTC researchers have developed several conven- tions for standardizing result reporting (Fenlon et al. 1999; Dachman and Zalis 2004). Results are reported on a per polyp as well as a per patient basis.
Results based on per poly analysis are the most rigor- ous as they imply direct comparison to colonoscopy.
Per patient analysis, however, is the more clinically relevant parameter in terms of referral of the patient for therapeutic optical colonoscopy. Polyp detection rates are usually grouped according to size as under 6 mm in diameter, 6–9 mm and 10 mm or greater.
Polyp location is given either by anatomical colon segment (six or eight segments) or by recording the linear colon center line distance from the anal verge from workstation software. Polyp measurements are usually given as the longest linear dimension either by 2D or 3D viewing and there is also an emerg- ing consensus that some form of a confidence limit modifier may be of some clinical merit.
To date, the most important published result has been that of a multi-center trial of screening in asymptomatic adults from the U.S. Department of Defense. (Pickhardt et al. 2003). In that pro- spective study of 1233 patients, CTC detected 96%
of polyps 8 mm or greater, was more accurate than optical colonoscopy, and found 55 polyps and one of 2 cancers missed by optical colonoscopy. CTC also gave a negative predictive value of 98% for any polyp greater than 10 mm in size, and showed that over 50% of patients had no polyps whatsoever pres- ent in their colon. The excellent results in that study were considered to be multi-factorial in nature and included the use of primary endoluminal 3D view- ing, aggressive double dose phosphosoda bowel preparation, knowledgeable radiologist readers, and the use of a novel segmental unblinding technique which produced a new consensus ‘ground truth’ by direct virtual and optical colonoscopic correlation.
As a result of that study, gastroenterologists in the U.S. and their national professional organizations conceded that CTC was a technique that was likely to be of wide value and encouraged its “use” by gas- troenterologists. However, a subsequent smaller study conducted by U.S. gastroenterologists several months later gave much poorer results (Cotton et al. 2004). However, that study was widely discredited by CTC radiologist researchers because of outmoded techniques and flawed study execution (Ferrucci et al. 2004; Pickhardt 2004; Halligan et al. 2004).
Nevertheless, some doubt as to the generalizability
of CTC performance was raised and the issue was left open as to whether or not additional studies of CTC in screening populations were really required.
Two such large multi-center trials are underway as of this writing, one in the U.S conducted by the American College of Radiology Imaging Network (ACRIN) and another in the U.K. carried out for the National Health Service by the Special Interest Group for Gastrointestinal and Abdominal Radi- ology (SIGGAR). However, the results of these two trials are not likely to be widely available before 2006–2007. In the meantime, rapid further tech- nical advances in CTC including the use of newer 16–64 slice multi-row detector scanners, laxative free colon cleansing schemes, and computer aided detection will become more widespread. Thus even these studies now well underway are destined to be characterized as outdated by the time their results are eventually published.
5
Acceptance of CTC
Several factors will likely converge in the very near future to precipitate wide-spread acceptance of CTC for colorectal screening.
• Continued excellent results from single center trials will add to existing cumulative data of its effi cacy.
• Evidence-based comparisons with existing approved screening tests (FOBT, fl exible sigmoid- oscopy, double contrast barium enema, colonos- copy) will show that none is perfect and that CTC has a suffi cient number of benefi ts and unique attractions to make it a legitimate addition.
• Continued deployment of modern CT scanner capability will make it apparent to community hospitals and private practitioners that there is a very low entry cost to introduce this new imaging product and hospital administrators will persuade radiologists, referring clinicians and gastroenter- ologists to adopt the procedure.
• Radiology professional organizations will develop
and promulgate appropriate practice standards,
guidelines, reporting schema, training curricula
and accreditation programs to insure that the
technique and the reading radiologists have cred-
ibility with the public and insurance carriers, as
well as other physicians and governmental policy
makers. The precedent of mammography in the
U.S. where quality standards have been developed
by the American College of Radiology has been invaluable to the public and radiologists alike.
Similar programs are necessary and in fact inevi- table for CTC.
• Gastroenterologists performing colonoscopy will increasingly understand that CTC is not so much a competitive threat, as it is in fact a case multi- plier in that it will detect and deliver patients to them with truly actionable large polyps for ben- efi cial polypectomy.
• Practical models for integration of virtual and optical colonoscopy into practice will evolve, especially those involving single day one stop shopping cross referral between radiologists and colonoscopists for positive fi ndings at CTC and by the same token for failed colonoscopies allowing an immediate follow-up same day virtual exam with only a single colon prep.
6
Conclusion
It can be predicted that final certifying examina- tions for young radiologists completing residency training will soon contain examples of CTC studies.
This is especially likely in as much as the current generation of trainees is no longer able to perform an adequate number of barium enemas to become proficient. Even though most diagnostic radi- ologists already consider themselves significantly overworked, it is now inevitable that CTC will be incorporated into daily radiologic practice in the very near future. Also predictable is that the clini- cal techniques for performing CTC will evolve even further with acquisitions faster, computer aided detection ubiquitous, small polyps ignored and rec- ommendations for further management conveyed in a structured and illustrated computer generated report. Indeed, in leading academic CT research departments, the future is already here.
References
Advisory Board (2004) www.advisoryboardcompany.com. The Advisory Board, Washington, DC
Bartrum C (2000) Personal communication
Bond JH (2001) Clinical relevance of the small colorectal polyp. Endoscopy 33:454–457
Coin CG, Bond WC, Stafford TO (1995) Computed tomographic colonoscopy. United States Patent 5,458,111. Oct 17, 1995
Cotton PB, Durkalski VL, Pineau BC et al. (2004) Computed tomographic colonography (virtual colonoscopy): a mul- ticenter comparison with standard colonoscopy for detec- tion of colorectal neoplasia. JAMA 291:1713–1719 Dachman AH, Zalis ME (2004) Quality and consistency in CT
colonography and research reporting. Radiology 230:319–
323
Edwards JT, Mendelson RM, Fritschi L, Foster NM, Wood C, Murray D, Forbes GM (2004) Colorectal neoplasia screen- ing with CT colonography in average-risk asymptomatic patients: community-based study. Radiology 230:459–464 Fenlon HM, Nunes DP, Schroy PC et al. (1999) A comparison
of virtual colonoscopy and conventional colonoscopy for the detection of colorectal polyps. N Engl J Med 341:1496–
1503
Ferrucci JT et al. (2004) Virtual colonoscopy. Letter to the Editor. JAMA 292:431–432
Halligan S et al. (2004) Virtual colonoscopy Letter to the Editor.
JAMA 292:431–432
Hara AK, Johnson CD, Reed JE et al. (1996) Detection of colorectal polyps by computed tomographic colography:
feasibility of a novel technique. Gastroenterology 110:284–
290
Iannacone R, Laghi A, Catalano C et al. (2003) Performance of lower dose multi-detector row helical CT colonography compared with conventional colonoscopy in the detection of colorectal lesions. Radiology 229:775–781
Leiberman DA, Weiss DG, Bond JH et al. (2000) Use of colonos- copy to screen asymptomatic adults for colorectal cancer.
N Engl J Med 343:162–174
Macari M, Bini EJ, Xue X, Milano A, Katx SS, Resnick D, Chanda- rana H, Krinsky G, Klingenbeck K, Marshall CH, Megibow AJ (2002) Colorectal neoplasms: prospective comparison of thin-section low-dose multidetector row CT colonogra- phy and conventional colonoscopy for detection. Radiology 224:383–392
Muto T, Bussey HJR, Morson BC (1975) The evolution of cancer of the colon and rectum. Cancer 36:2251–2270
Pickhardt P (2004) Virtual colonoscopy. Letter to the Editor.
JAMA 292:431
Pickhardt PJ, Choi JR, Hwang I et al. (2003) CT virtual colonos- copy to screen for colorectal neosplasia in asymptomatic adults. N Engl J Med 349:2189–2198
Royster AP, Fenlon HM, Clarke PD et al. (1997) CT colonos- copy of colorectal neoplasms; two-dimensional and three- dimensional virtual-reality techniques with colonoscopic coreelation. AJR 169:1237–1242
Svensson MH, Svensson E, Lasson A et al. (2002) Patient acceptance of CT colongraphy and conventional colo- noscopy: prospective comparative study in patients with or suspected of having colorectal disease. Radiology 222:337–345
Vining DJ, Gelfand DW (1994) Non-invasive colonoscopy using helical CT scanning. 3D reconstruction and virtual reality.
Syllabus. 23rd Annual Meeting Society of Gastrointestinal Radiologists, Maui, Hawaii
Winawer SJ, Zauber AG (2002) The advanced adenoma as the primary target of screening. Gastrointestinal Endoscopy Clinics of North America, vol 12, pp 1–9
Yee J, Akerkar GA, Hung RK, Steinauer-Gebauer AM, Wall SD, McQuaid KR (2001) Colorectal neoplasia: performance characteristics of CT colonography for detection in 300 patients. Radiology 219:685–692
